Results
External regulation
There were two significant multivariate effects of time observed for external regulation. Whereas there is apparently some changes in the external regulation scores over time in the four studies, they were not all significant. One significant effect was for Study 3, reported in Table 1. Follow up univariate tests showed the significant change (decrease) to occur between Time 2 (5 weeks) and Time 3 (10 weeks). The other significant change was in Study 5, reported in Table 1. Follow up univariate tests showed the significant change (decrease) to be from Time 1 to Time 2 (6 weeks), F(1,62) = 5.51, p < .05, η2 = .08.
Introjected regulation
Only one significant multivariate main effect for time was observed for time for introjected regulation. This was for Study 4 (reported in Table 1). Follow up univariate tests revealed that introjected regulation increased from baseline to 8 weeks (mid-point for Study 4), F(1,133) = 7.68, p < .006, η2 = .06, but there was no significant change from week 8 to week 16 (F(1,133) = .25, p = .06). The rest of the studies showed similar trends for small increases in introjected regulation, but none achieved statistical significance.
Identified regulation
Significant main effects of time were observed for identified regulation in Studies 3, 4, and 6, and these are reported in Table 1. For Study 3, follow up repeated measures ANOVA revealed a significant increase from baseline to 8 weeks, F(1,76) = 8.43, p < .005, η2 = .10, and a significant increase from 8 weeks to 12 weeks, F(1,76) = 20.96, p < .0001, η2 = .22. Similarly, in Study 4, a significant increase was observed from baseline to week 8, F(1,133) = 59.45, p < .0001, η2 = .31, but not from week 8 to week 16, F(1,133) = .04, p = .84. For Study 6, again an increase was observed from baseline to 12 weeks, F(1,83) = 13.15, p < .0001, η2 = .14, but not from week 12 to week 14, F(1,83) = .69, p = .41. The rest of the studies showed small increases over time that did not achieve statistical significance.
Intrinsic regulation
Finally, significant multivariate main effects of time were observed for intrinsic motivation in Studies 3 and 4. These are reported in Table 1. Follow up repeated measures ANOVA revealed no increase from baseline to 5 weeks in Study 3, F(1,76) = 1.02, p = .32, but a significant increase from week 5 to week 10, F(1,76) = 56.68, p < .0001, η2 = .43. In Study 4, a significant increase was observed from baseline to 8 weeks, F(1,133) = 59.45, p < .0001, η2 = .31, but not from week 8 to week 16, F(1,133) = .04, p = .84. The rest of the comparisons showed small increases over time that did not achieve statistical significance.
Differences between regular exercisers and initiates after exercise
Examination of the means for the regular exercisers compared to the initiates showed that their scores were considerably higher for the self-determined forms of regulation (identified and intrinsic) at all time points, even the last observed scores of the initiates at the end of each exercise program.
The values of d are presented in Table 3. According to Cohen (1992) values of d around 0.2 are small effects, around 0.5 are medium effects and around 0.8 are large effects. An average of the ds comparing each of the four groups of initiate exercisers to the Study 1 regular exercisers and then to the Study 2 regular exercisers was calculated for each form of regulation. These are also reported in Table 3. The average ds for external regulation suggested a large effect compared to Study 1 and a small effect compared to Study 2. The average ds for introjected regulation also revealed a large and small effect compared to Studies 1 and 2 respectively. The average ds for identified regulation were large and medium. Finally, for intrinsic regulation the average ds both showed large effects. Thus, the magnitudes of the differences between initiate exercisers and regular exercisers, even after six weeks to six months of exercise are meaningful for external, identified and intrinsic regulations, such that exercise initiates remain (on average) lower on all types of regulation.
Table 3.
Values of Cohen’s d values for comparisons of initiate exercisers (Studies 3 through 6) to regular exercisers (Studies 1 and 2).
Form of regulation Comparison study Study 3 initiates Study 4 initiates Study 5 initiates Study 6 initiates Average d across 4 studies M (SD)
External Study 1 0.95 1.20 1.90 1.70 1.47 (.39)
Study 2 0.00 0.25 0.63 0.51 .34 (.28)
Introjected Study 1 0.84 0.93 0.68 1.20 .91 (.22)
Study 2 0.16 0.29 0.10 0.47 .25 (.16)
Identified Study 1 2.25 2.09 2.22 1.93 2.12 (.15)
Study 2 0.14 0.35 0.69 0.34 .38 (.23)
Intrinsic Study 1 2.28 2.26 2.75 2.52 2.45 (.23)
Study 2 0.51 0.68 1.01 0.80 .75 (.21)
